RSX aren't capable of HDR+FSAA?

[Xenus]

I could of swore I read somewhwere the Xenos only supported MSAA on FP10 and not on FP16.

 

[expletive]

Is FP10 a Xenos-specific mode or is it just what's supported in 'hardware' on Xenos?

 

[Titanio]

I can comment on Xenos since I haven't signed any NDA with Microsoft: Xenos supports FP16 HDR with MSAA, but it can't automatically resolve the frame buffer whilst dumping it out to the external memory. One needs to download that tile into main ram and to resolve it using a pixel shader.

That's interesting to know. I thought blending on FP16 wasn't possible? Perhaps that's a disjoint issue.

Unfortunately this is not true, the most noticeable tread off we made is trading ALU ops for memory storage and bandwidth.

Sure, but I meant in terms of dynamic range for your lighting vs FP16 - purely in terms of the HDR. It's roughly equivalent or better in some instances, yes?

edit - oops, missed your post just earlier. i'll take it as a yes

 

[expletive]

Unfortunately this is not true, the most noticeable tread off we made is trading ALU ops for memory storage and bandwidth.

I guess it was a design decision that HDR was more important than whatever those additional ALU ops would/could have provided? How would you chracterize the amount of ops your HDR implementation eats up?

 

[nAo]

That's interesting to know. I thought blending on FP16 wasn't possible? Perhaps that's a disjoint issue.

blending is not possible, but MSAA is possible.

Sure, but I meant in terms of dynamic range for your lighting vs FP16 - purely in terms of the HDR. It's roughly equivalent or better in some instances, yes?

Yes

I guess it was a design decision that HDR was more important than whatever those additional ALU ops would/could have provided? How would you chracterize the amount of ops your HDR implementation eats up?

RSX has a lot of computational units, there's nothing wrong to put bit more pressure on them
The amount of clock cycles per pixel can vary cause sometimes the hw can pair and/or co-issue more or less istructions. A typical cost is four 4 cycles, sometimes is 3..some other is 5 or 6. To be fair I did not bother to optmize it yet.

 

[Brimstone]

What does options available in PC software have to do with RSX? The incentive may not be there to implement a solution for a select few nVidia cards in the PC space, but in a closed box, you can tune your approach to one set of hardware you know everybody has.

Microsoft and ATI optomized their 360 hardware to achieve full throtle performance given the limitations of the console enviroment. It's one of the reasons why the 360 has 10mb of eDRAM with the ROPS and other circutry combined to allow for A.A. + HDR without slowing down the rest of the GPU.

At the end of the day, programmers aren't going to be able to create more bandwidth and processing power than a system has.

 

[expletive]

RSX has a lot of computational units, there's nothing wrong to put bit more pressure on them

Give 'em hell!

The amount of clock cycles per pixel can vary cause sometimes the hw can pair and/or co-issue more or less istructions. A typical cost is four 4 cycles, sometimes is 3..some other is 5 or 6. To be fair I did not bother to optmize it yet.

For the uninitiated, is that: a lot, a fair amount, not much, or inconsequential?

 

[nAo]

For the uninitiated, is that: a lot, a fair amount, not much, or inconsequential?

Appplying it on every pixel of a 720p image would take 2%-3% of a single frame time (60 fps),
this means that most of the time it's well worth its cost

 

[Titanio]

Microsoft and ATI optomized their 360 hardware to achieve full throtle performance given the limitations of the console enviroment. It's one of the reasons why the 360 has 10mb of eDRAM with the ROPS and other circutry combined to allow for A.A. + HDR without slowing down the rest of the GPU.

At the end of the day, programmers aren't going to be able to create more bandwidth and processing power than a system has.

I don't see how this answers my query. The status of PC software with HDR & MS/FSAA being either/or need not apply with PS3 games necessarily, and whether you can do MSAA or not doesn't affect your dynamic range, which seemed to be your concern.

Programmers can't magic up more bandwidth, but depending on requirements, they can indeed get dynamic range better than the hardware-supported solution you hold out, with MSAA, on PS3 Even using the 'easy route' of a FP16 buffer yields you excellent HDR - if that is indeed your concern - whatever about AA.

Appplying it on every pixel of a 720p image would take 2%-3% of a single frame time (60 fps)

Hehe, that's what I came up with (well, 1.666*% for a 4-cycle cost)..I assumed the cost was once per pixel per frame? Rather than with every write to the pixel?

 

[scooby_dooby]

AFAIK KZ renderer is fairly different from HS renderer and makes perfctly sense for them to not use our color space, but it's not an issue related to lighting requirements, in fact we support a VERY broad (and accurate!) range , thousand times broader than FP10 on Xenos.

So the renderer is what determines whether or not NAO32 is a viable alternative? Can you explain that a little further?

 

[nAo]

Hehe, that's what I came up with (well, 1.666*% for a 4-cycle cost)..I assumed the cost was once per pixel per frame? Rather than with every write to the pixel?

You're right, but I made my estimation not using the final clock..emh..

 

[nAo]

So the renderer is what determines whether or not NAO32 is a viable alternative? Can you explain that a little further?

Unfortunately in this case I can't/I do not want to explain it, sorry

 

[Titanio]

You're right, but I made my estimation not using the final clock..emh..

Neat, thank you

 

[Brimstone]

I don't see how this answers my query. The status of PC software with HDR & MS/FSAA being either/or need not apply with PS3 games necessarily, and whether you can do MSAA or not doesn't affect your dynamic range, which seemed to be your concern.

FSAA + HDR is one more thing that PS3 developers will have to devote more time on to get working (jump thru hoops), while XB360 developers aren't going to sweat the details as much.

The reality is FSAA + HDR on the 360 is going to be quicksilver fast regardless of game style and should allow for other cool effects to be running at the same time.

 

[Mmmkay]

Okay after looking at that higher quality Warhawk footage, it appears to have HDR. During the speech they mention that it uses 4xAA. P.s. Clouds, oh my the clouds!

I wonder if nAo32 has been SCE WWS's village bicycle?

 

[Mordecaii]

FSAA + HDR is one more thing that PS3 developers will have to devote more time on to get working (jump thru hoops), while XB360 developers aren't going to sweat the details as much.

The reality is FSAA + HDR on the 360 is going to be quicksilver fast regardless of game style and should allow for other cool effects to be running at the same time.

Just because they can implement it faster doesn't mean it'll be comparable quality. Did you see the thread earlier stating how much better NAO32 was than FP10? And I believe it was mentioned earlier that FP16, which would give close to the same quality as NAO32, would be difficult on the 360 do to the 10mb EDRAM... So there are always tradeoffs, and while FP10 plus MSAA will be possible and probably easier to implement, it will be lower quality.

 

[mckmas8808]

Okay after looking at that higher quality Warhawk footage, it appears to have HDR. During the speech they mention that it uses 4xAA. P.s. Clouds, oh my the clouds!

I wonder if nAo32 has been SCE WWS's village bicycle?

No seriously I would really like to know if nAo32 is really being used in SCE WWS? Because HDR and 4x AA on a launch game is great for a GPU that's said not to be able to do it.

 

[overclocked]

I don't see how this answers my query. The status of PC software with HDR & MS/FSAA being either/or need not apply with PS3 games necessarily, and whether you can do MSAA or not doesn't affect your dynamic range, which seemed to be your concern.

Programmers can't magic up more bandwidth, but depending on requirements, they can indeed get dynamic range better than the hardware-supported solution you hold out, with MSAA, on PS3 Even using the 'easy route' of a FP16 buffer yields you excellent HDR - if that is indeed your concern - whatever about AA.



Hehe, that's what I came up with (well, 1.666*% for a 4-cycle cost)..I assumed the cost was once per pixel per frame? Rather than with every write to the pixel?

What formula did you use, i got about ~2,5% but of course i as you counted the clockspeed *we* been told by Sony.
btw i took the average of the cycles but i see you only counted with a 4cycle pressumption.

 

[nAo]

4 clock cycles * 60 (fps) / 550.000.000 Hz / 24 * 921600 pixels * 100 = 1.67%

 

[overclocked]

Wow!

4 clock cycles * 60 (fps) / 550.000.000 Hz / 24 * 921600 pixels * 100 = 1.67%

What i thought, i just tooked the "worst" case scenario or how we say it with more clock cycles.
I must say im impressed with what you have done nAo and im looking forward to see the stuff on E3, perhaps you have optimized it then also hehe!

No seriously I would really like to know if nAo32 is really being used in SCE WWS? Because HDR and 4x AA on a launch game is great for a GPU that's said not to be able to do it.

Yeah would be great too know. nAo hasnt said anything besides that KZ didnt.

I was impressed by the Warhawk footage and that he mentioned just 4AA aswell as the surficematerials on the capital ships etz, i just want Lair with the others so bad!